Champa implements adaptive request-rate control: when the AMP cache returns an unexpected response (e.g., HTTP 302), the client halves its polling rate. Once the polling interval exceeds one second the client is considered fully rate-limited and must restart after a multi-hour cooldown. This mechanism keeps the client from burning through its allowance when cache responses degrade.

Champa's AMP-cache tunnel delivers at most ~15 KB/s throughput before AMP rate-limiting activates. Sessions last approximately 15 minutes before the rate limit triggers; after that, users must stop the client and wait several hours before resuming.

Issue body (rate-limit description) evaluation

As of June 2025, Iran blocks GitHub.com and Telegram but leaves the IP address of google.com reachable and Google Drive accessible. This selective blocking pattern created the operational window for Champa: because google.com is reachable, its AMP cache (cdn.ampproject.org) can be reached by fronting on that IP, then used to proxy connections to otherwise-blocked destinations.

Issue body (opening description) deployment

Because GitHub is blocked in Iran, Champa distributes binaries through three alternative channels: a Telegram bot (@getchampabot), Google Drive folders, and curl commands documented in the GitHub issue thread itself. The issue thread remains accessible via Google cache / AMP even when GitHub's domain is blocked, creating a bootstrapping loop.

Download section (§1) deployment

Champa routes connections by specifying an AMP cache URL (https://cdn.ampproject.org/) with google.com as the front domain, then uses the SOCKS5 proxy interface locally. The server's public key is hardcoded in the connection string, providing a fixed authentication token that prevents trivial impersonation of the proxy by a censor injecting a fake server response during active probing.

Setup section (§2) defense

During the June 2025 Iran shutdown, circumvention tool performance diverged sharply by transport design. Psiphon's multi-protocol architecture sustained 1.5 million concurrent users—roughly one-third of its normal Iranian base. Lantern's "proxyless" protocol (domain-fronting via CDN, ~40% of Lantern's Iranian traffic) showed moderate success. Tor usage collapsed during the blackout but bridge connections surged and rebounded quickly after lifting. BeePass (serving 500k+ daily users at shutdown onset) used live A/B testing of port/obfuscation-prefix combinations to probe the censors' blocking parameters in real time. The Ceno Browser's P2P network grew from 600 active peers on June 13 to ~8,000 by July 11, indicating that decentralized fallback paths stayed up even during peak blocking.

2025-iran-shutdown-measurement §Tool Performance dpithrottling

In April 2026, Russia nearly completely blocked Telegram, reducing availability without VPN to 5% during the first ten days; Russia's VPN user base was estimated at no less than 40% of the population. Beginning in mid-April, over 20 major Russian websites and platforms (including Gosuslugi, Ozon, Yandex, Sberbank, VK) began restricting access for VPN users, and Russia introduced a whitelist-based service model permitting only state-approved services under certain network restriction states.

2026-anon-2026-27-vpn §4 (俄罗斯 Telegram / OSW section) ip-blockingdpithrottling

After the shutdown lifted (Scan 3), DNS-over-UDP degraded severely — 89.18% of the 9,000 queried domains were affected — but the paper attributes this to general UDP transport instability rather than domain-specific censorship, corroborated by GitHub user reports of ongoing UDP issues. TCP-based protocols (HTTP, TLS) returned to pre-shutdown censorship levels (~15%) in Scan 3, confirming the post-shutdown persistence effect was UDP-specific and not a broadening of the blocklist.

2026-anon-insights-into-iranian §2.2.4, Table 1 throttlinghttp3-quic-block

Iran's June 2025 shutdown was enforced using fine-grained, service-level techniques rather than BGP withdrawals or a single 'big switch' event: individual protocols were selectively blocked before and after the 7-day full blackout (June 18–25), with the total preparation and recovery periods each lasting roughly 4–5 days. Cloudflare Radar showed HTTP traffic reaching near-zero by June 18 but partial recovery already on June 21, while the authors' vantage point (AS57497) remained unreachable until June 25, confirming heterogeneous recovery across Iranian networks.

2026-anon-insights-into-iranian §2, §3 dpihttp3-quic-blockthrottling

Article 19 documents that Iran's National Information Network (NIN / SHOMA) was designed with explicit reference to China's Great Firewall as a model, with institutional mirroring: Iran's Supreme Council of Cyberspace parallels China's Cyberspace Administration of China, and both governments share a "cyber sovereignty" doctrine used to justify domestic content controls and cross-border technology transfer. The report frames Iran's filtering infrastructure as deliberately architected to replicate GFW capabilities, not as an independently developed system.

2026-article19-tightening-the-net §2, §3 dpisni-blockingdns-poisoningip-blockingbgp-hijackthrottling

Article 19 documents that Iran combines technical filtering with formal coercion of major foreign platforms (including Telegram, Instagram, and WhatsApp) to comply with content removal orders under threat of full blocking. The report notes that Iran's 2022 Women Life Freedom protests accelerated platform blocking when foreign operators refused compliance, demonstrating that the censorship system operates in two modes: coerce-and-allow for compliant platforms, block for non-compliant ones. Domain fronting via these platforms is therefore subject to sudden revocation if political conditions change.

2026-article19-tightening-the-net §6 ip-blockingdns-poisoningthrottling